Upgrading the Semantics of the Relational Model for Rich Owl 2 Ontology Learning

This paper is interested in the ontology learning from relational databases (RDB) that exploits already approved semantics about a domain and translates them to application ontology. However, the relational model is recognized to be less expressive and incapable to support some conceptualizations. Without an explicit model of the domain semantics in the relational model, the automatic learning of ontology risks to infer incorrect semantics. In this paper, we give some proof case studies and we propose a model to upgrade the semantics of the relational model, before the ontology learning. The paper presents the constructs of the proposed model and it shows how they are translated to constructs of OWL 2 ontology.